|
|
| i4004 |
| Posted: Dec 10 2003, 09:54 PM |
 |
|
Advanced Member
Group: Members
Posts: 2432
Member No.: 4935
Joined: 24-June 03
|
pages are cca.500kB or less....
http://free-zd.hinet.hr/ikostic3/laced.htm
[edit;pages mirrored to this web-space,as 0catch was broken because of "excessive"(20MB per day) usage....]
[edit2;ffvfw page updated with
"ffvfw versions tips"
considerable speed-up can be achieved with proper ffvfw version....]
--------------------
my signature:
 |
|
| BaronVlad |
| Posted: Dec 10 2003, 11:29 PM |
|
|
Terrorhörnchen
Group: Moderators
Posts: 335
Member No.: 28
Joined: 12-July 02
|
Hmmmm... sorry, page seems to be off ATM:
| QUOTE | 20 Meg Limit Exceeded!
This site has exceeded its limit of 20 MB of transfer per day. The account can be upgraded to a paid account to increase the transfer limit to 3 Gigs of transfer per month in your 0catch.com user tools section. |
--------------------
Hochachtungsvoll
BaronVlad
Deutsch >> Capture FAQ - Capture Guide
Englisch >> Capture FAQ - Capture Guide |
|
| i4004 |
| Posted: Dec 11 2003, 05:35 AM |
|
|
Advanced Member
Group: Members
Posts: 2432
Member No.: 4935
Joined: 24-June 03
|
oups!
my mistake,i linked it to arstechnica forum too...that was too much as it seems....or it's just the hoster wanting my money badly....heh
take it easy people,one by one...(i know test indeed is interesting one...  )
i'll try to find a better,unlimited web-space tomorrow(or make few of these lousy free-web mirrors),and copy it there too....
thanks for pointing it out!
i will solve it......
--------------------
my signature:
|
|
| Pamel |
| Posted: Dec 11 2003, 07:48 PM |
|
|
matroska team member
Group: Moderators
Posts: 151
Member No.: 1353
Joined: 15-December 02
|
It would help if you just made a rar of the webpage/graphics and then seeded with a Bittorrent link.
--------------------
Please take anything I say with many many grains of salt. I can't tell the difference between my AVI and a hole in the ground. |
|
| i4004 |
| Posted: Dec 11 2003, 11:09 PM |
|
|
Advanced Member
Group: Members
Posts: 2432
Member No.: 4935
Joined: 24-June 03
|
i believe that bittorrent option implies that a user is fluent in bittorrent,right?
so i wouldn't call it broadly excepted.....[anyhow,i will appreciate your comment...PM me with few short explanations if that can be done in a simple,universal way(for example;people generally tend to hate command line or many complications..heck,kazaa is too much for some..).....i myself didn't got to usage part of bittorrent anyhow....anyway,send your comment and i'll look into it....]
btw.why rar and not 7zip?
the page content is now mirrored here;
http://free-zd.hinet.hr/ikostic3/laced.htm
if anyone has problems now let me know....
[edit:i just looked at the dloading side of bittorrent and it's easy enough.... i must see if uploading will work as good..]
[edit2:as i see it now,i would have to seed it ,but in sync with few dloaders....so i would have to give a time window in which you would be sucking my stuff......we'll see...i think i can aquire another unlimited bandwidth 5MB web-space from a friend...]
--------------------
my signature:
|
|
| Morsa |
| Posted: Dec 12 2003, 04:29 AM |
|
|
Moderator of the Vdub support board
Group: Moderators
Posts: 640
Member No.: 246
Joined: 9-September 02
|
Amazing results!!
Good Job Ivo!
I just wonder if recording original material onto a DVD with Mpeg4 and seeing it in a Divx player would work OK or if any difference in quality would be noticeable. |
|
| fccHandler |
| Posted: Dec 12 2003, 05:21 AM |
|
|
Administrator n00b
Group: Moderators
Posts: 3961
Member No.: 280
Joined: 13-September 02
|
Well, you've proven that MPEG-4 looks better at low bitrates (and 2Mbit/s is way too low for MPEG-2 at DVD resolution). But I suspect that if you compared both codecs at 8Mbit/s, the MPEG-2 video would be a better reproduction of the original source than the MPEG-4 video. (Whether your eyes could detect it is another matter...)
--------------------
May the FOURCC be with you... |
|
| i4004 |
| Posted: Dec 12 2003, 02:21 PM |
|
|
Advanced Member
Group: Members
Posts: 2432
Member No.: 4935
Joined: 24-June 03
|
@morsa;i dunno about laced mpeg4 playback support on standalone divx players....i also feel that 720x576 constrain (on such players,if i heard ok informations) is not so good;some of my stuff gets to be 768x576 too....(although rarely....)
quality?i have watched all of this on my interlaced tv-out,and mpeg4 looks MIGHTY fine!
perhaps you should ask on doom9 hardware forum about that..i myself am not so hot for hardware players of any sort...
but i believe interlacing support is crucial,as we're not getting rid of interlaced scan any time soon....that's for sure....(and even hdtv has an interlaced mode:1080i)...there's sports there's music videos there's live-tv there's...huh...porn (  )
@fcc; 8Mbit/s only?why not go to 15Mbit/s? ( )
that way i may put perhaps 2hrs of video on upcoming aod disc....
seriously,with these releases of ffvfw and xvid,i'm really leaving mpeg2 behind me (i hope...first test is *very* encouraging....next one will be "Outkast-Hey Ya" music video...a vhs rip to-do(as i can smell it even on watching vhs on tv that it has lots of blends....) )
on the matter of eyes detecting the stuff;well,you said it best..(hehe).....
also,i get your thoughts.as on my 1,5Mbit progressive test mpeg1/2 came sharpest.....but it's motion estimation wasn't so shiny....
as i said in my test;
"but hold your horses,there's more to video compression than sharpness (in which mpeg1/2 is a clear winner,as illustrated in above gif image......mpeg1/2 wins sharpness test with ease...) :
mpeg1/2 ME algo is usually visibly worse than that of mpeg4,and the lower you go with the bitrate,the more obvious this becomes.....i have also done 448x336 @770kbit/s and compared mpeg1/2 to mpeg4.....no doubt,ME performance of mpeg1/2 deteriorates much more rapidly than mpeg4's.....although spatial performance(sharpness of stiller scenes) stays excellent with mpeg1/2 even on that low bitrates..."
but one can easily enough change ffvfw and xvid's quantization matrix so better sharpness is achieved...(if need be....on 7xx x 576 i believe that shouldn't be necessary anyhow...)
btw.do you know why mpeg4 has better temporal compression(more efficient) than mpeg2?
is it storing MV's with higher compression than mpeg2?is it finding better matches?
please put in a few words if you have any knowledge on that...(to save me some time on reading the standard specs....  )
--------------------
my signature:
|
|
| Anpu |
| Posted: Dec 13 2003, 01:21 PM |
|
|
Unregistered
|
| QUOTE (i4004 @ Dec 12 2003, 08:21 AM) | btw.do you know why mpeg4 has better temporal compression(more efficient) than mpeg2?
is it storing MV's with higher compression than mpeg2?is it finding better matches?
|
MPEG2 uses one motion vector per macroblock (16x16).
MPEG4 allows 1 or 4 (one for each 8x8 part) vectors per macroblock, which are differentially coded using vectors of three neighboring blocks.
When bit-rate is high enough MPEG4 can have more detailed motion info than MPEG2 with the same bit budget.
At low bit-rate MPEG4 can store the same motion info as MPEG2 but more compactly, so it will have more bits left for the DCT coefficients. |
|
| fccHandler |
| Posted: Dec 13 2003, 06:00 PM |
|
|
Administrator n00b
Group: Moderators
Posts: 3961
Member No.: 280
Joined: 13-September 02
|
| QUOTE (Anpu @ Dec 13 2003, 09:21 AM) | | MPEG2 uses one motion vector per macroblock (16x16). |
Not necessarily. MPEG-2 can use different vectors for each field of an interlaced macroblock. It can also use different vectors for the upper and lower (16x8) halves of a non-interlaced field picture, or "dual prime" prediction (which I don't fully understand). Of course, it's up to the encoder to support all these modes...
--------------------
May the FOURCC be with you... |
|
| Anpu |
| Posted: Dec 13 2003, 09:13 PM |
|
|
Unregistered
|
| QUOTE (fccHandler @ Dec 13 2003, 12:00 PM) | | MPEG-2 can use different vectors for each field of an interlaced macroblock. |
True. But since MPEG-2 does not use inter-block prediction for motion vectors, for 16x8 blocks to be useful rather high bitrate is necessary. Otherwise there will be not enough bits left to encode the residuals.
Shortly put - for interlaced video MPEG2 needs even higher bitrate to get motion not as good as MPEG-4. |
|
| i4004 |
| Posted: Dec 13 2003, 09:25 PM |
|
|
Advanced Member
Group: Members
Posts: 2432
Member No.: 4935
Joined: 24-June 03
|
ok,thanks to both of you;
fcc,don't get my words wrong;i'm not doing any silly talk ala dxn guys (like "divx5 is 10x more effective compression than mpeg2" etc.),and also i have stated how there's NO truth in a statement that "mpeg4 is twice as good as mpeg1/2"..(dunno if it was in english or on my croatian NG).....
that's crap!
differences between mpeg2 and mpeg4 are not that great at all.....and on higher bitrates i indeed can say that mpeg1/2 looks mighty fine....
i have also done my part to correct these misconceptions about mpeg's in my encoding tests (as who else included mpeg1 head to head to mpeg4.....and again,differences were not that big at all...)
tmpgenc makes wonderfull mpeg stuff.....cce too but on slightly higher bitrates than tmpgenc
but if it comes to making 2Mbit/s or 3Mbit/s i'll pick 2Mbit stuff if it offers same quality as that other compression on 3Mbit....(i just did 768x576@1,7Mbit/s ffvfw laced.....lookin fine(..just watched it on my tv-out...)....i couldn't do that with present top-notch mpeg2 encoders...(i could ,but with lesser quality) )
btw. i have found these words mentioned in the context:
"If you can ingest that, the next step is really getting to know the differences between MPEG-2 and MPEG-4, such as inter4v mode, new prediction modes, etc.."
[by -h on doom9]
i guess inter4v is what anpu just described as "MPEG4 allows 1 or 4 (one for each 8x8 part) vectors per macroblock,"
also i've found some stuff in h263 papers etc.
the answer anpu gave was kind of expected;effective MEstimation making some more room for dct's....
cheers
/ivo
--------------------
my signature:
|
|
| i4004 |
| Posted: Dec 15 2003, 09:49 AM |
|
|
Advanced Member
Group: Members
Posts: 2432
Member No.: 4935
Joined: 24-June 03
|
ffvfw page updated with
"ffvfw versions tips"
considerable speed-up can be achieved with proper ffvfw version....
--------------------
my signature:
|
|
| K0KE |
 Posted: Dec 18 2003, 10:51 PM |
|
|
Unregistered
|
A liitle bit off topic...
Lookin at those pictures, and bearing in mind your
attitude towards div3, I just couldnt resist it.
I compared a bit your mpeg4 tests and:
good: divx3
better: ffvfw
the best: xvid
I can explain and prove it if needed.
Regards
|
|
| i4004 |
| Posted: Dec 19 2003, 08:38 PM |
|
|
Advanced Member
Group: Members
Posts: 2432
Member No.: 4935
Joined: 24-June 03
|
swap the glasses:those aren't doing a good job for you (  )......
1-mpeg2=crap
2-divx3=crap(doesn't have lacing support..it's a progressive codec...acts funny on interlaced stuff)
3-xvid=ok
4-ffvfw=good.......(ie. better than xvid by a slight (but visible) margin...visible on fade-out sequence...)
sure,you can explain and prove by some screenshots;as stated on my web,everyone is welcomed to prove me wrong....(wether it be interlaced or progressive tests i did....)
i'm listening....
--------------------
my signature:
|
|
